BibTeX:

@article{IJIRSTV4I12005,
     title={An Effective Approach to Shadow Removal and Skin Segmentation using Deep Neural Networks},
     author={Berlin S John, Hari S and Anitha Rajam B S},
     journal={International Journal for Innovative Research in Science & Technology},
     volume={4},
     number={12},
     pages={35--37},
     year={},
     url={http://www.ijirst.org/articles/IJIRSTV4I12005.pdf},
     publisher={IJIRST (International Journal for Innovative Research in Science & Technology)},
}

One of the greatest challenge of traditional skin segmentation is false positive rate increases when skin tones are under shadow conditions and overlapping of skin pixels and non-skin pixels. This paper presents an effective approach to shadow removal and skin segmentation in both RGB and HSV color space and it is a deep neural networks to learn high-level representations of skin tones. Firstly, shadow from the input images or frames of video is detected by using hybrid shadow detection algorithm and then removed using correlation based neighbourhood matching algorithm. After the image enhancement, the skin segmentation is done by using triclass thresholding. Skin region recognition can be done by using VGGNet and has learned rich feature representation for a wide range of images. Instead of predicting each pixels individually, we utilize block of pixels for skin segmentation to avoid overlapping of skin pixels and non-skin pixels.

Pixels, Vggnet, Skin Segmentation, Correlation, Images